Stochastic Nodal Adequacy Pricing (SNAP) Platform: A Methodology for Dealing with Weather and Operational Uncertainty in Market Operations

Abstract

A probabilistic extension of electricity production cost minimization tools that supports the use of high-fidelity models is introduced. These tools introduced are able to accurately simulate the temporal and spatial relationships affecting system physics and economics. The ability to use high-fidelity models enables accurate calculation of dual variables and their use in defining reliability metrics that accurately represent the economic and engineering characteristics of all resources. In particular, the use of dual variables captures impacts of time-coupled resources and constraints such as storage and limited fuel supply. By bringing economic metrics directly into reliability analysis, we can supplement traditional reliability metrics with economically justified operations. These techniques and their computational performance are illustrated using a high-fidelity model of a real-sized US market, more specifically ERCOT (Electric Reliability Council of Texas). The model includes MIP based security-constrained unit commitment, realistic operational details, and co-optimization of energy and reserves.